Is osseous dysplasia a primary feature of neurofibromatosis 1 (NF1)?

Clin Genet 2005: 67: 378–390 Printed in Singapore. All rights reserved

Copyright # Blackwell Munksgaard 2005

CLINICAL GENETICS doi: 10.1111/j.1399-0004.2005.00410.x

Mini Review

Is osseous dysplasia a primary feature of neurofibromatosis 1 (NF1)? Alwan S, Tredwell SJ, Friedman JM. Is osseous dysplasia a primary feature of neurofibromatosis 1 (NF1)? Clin Genet 2005: 67: 378–390. # Blackwell Munksgaard, 2005 Characteristic skeletal lesions are a cardinal feature of the autosomal dominant condition, neurofibromatosis 1 (NF1). The most frequently involved skeletal sites are the sphenoid wing, vertebrae, and tibia. Osseous lesions may range in severity in NF1 but are often progressive. They may lead to serious clinical consequences and be resistant to treatment. The skeletal lesions of NF1 are usually considered to be ‘dysplasias’, i.e. primary defects of bone, although there is no direct evidence supporting this interpretation. Moreover, it is difficult to understand why a generalized dysplasia of bone would produce focal lesions that show such a striking predisposition to only a few bones. We review the clinical and pathological features of NF1 skeletal lesions and propose that they result from an abnormal response of NF1 halpoinsufficient bone to abnormal mechanical forces rather than from a primary osseous dysplasia.

S Alwana, SJ Tredwellb and JM Friedmana a Department of Medical Genetics, University of British Columbia, and b Department of Orthopedics, British Columbia’s Children’s Hospital, Vancouver, BC, Canada

Key words: bone – dysplasia – neurofibromatosis 1 – pathogenesis – sphenoid wing – tibra – vertebrate Corresponding author: Sura Alwan, Department of Medical Genetics, Room 300H Wesbrook Building, 6174 University Boulevard, Vancouver, BC V6T 1Z3, Canada. Tel.: þ1 604 822 2749; fax: þ1 604 822 5348; e-mail: [email protected] Received 27 August 2004, revised and accepted for publication 9 December 2004

Neurofibromatosis 1 (NF1) is one of the most common autosomal dominant disorders, occurring with an incidence of around one in 3500 (1, 2). The most frequent characteristics include cafe´au-lait spots, axillary freckling, iris Lisch nodules, and multiple neurofibromas, which may present as discrete cutaneous lesions or deeper plexiform tumors involving the peripheral and spinal nerves. The average age at death among people with NF1 is about 15 years earlier than the population norm (3), with malignancies (most often malignant peripheral nerve sheath tumors or brain tumors) (4, 5) and cardiovascular disease being the most common causes of early death. The penetrance of NF1 is complete, but the manifestations are extremely variable, although they generally increase with age (6, 7). Many of the features of NF1 involve neural crest-derived cells, and the condition has consequently been classified as a neurocristopathy (8). However, it is now clear that NF1 also involves nervous, vascular, and skeletal system components that have no apparent relationship to the neural crest. Children with NF1 characteristically 378

have mild shortness of stature and macrocephaly (9–11). Decreased growth in NF1 usually affects the whole skeleton proportionately (10, 12). Macrocephaly, on the other hand, usually results from enlargement of the brain rather than overgrowth of the skull in people with NF1 (12). Brooks and Leham (13) published the first classification of the osseous changes of NF1 in 1924. Following that, many authors have described orthopaedic manifestations that are characteristic of NF1 and have estimated the prevalence of such changes to be in the range of 30–70% (14–18). A variety of skeletal lesions may occur in NF1 patients, including localized overgrowth, erosion of bone by an adjacent neurofibroma, spinal curvature that resembles common adolescent scoliosis, and certain unusual osseous lesions that are so characteristic of NF1 that they constitute one of the standard diagnostic criteria (1). As with other NF1 clinical features, the orthopaedic manifestations can range from mild to severe. The characteristic bony lesions of NF1 are generally considered to be dysplasias (12, 16). We believe that in order to understand the

Osseous dysplasia

pathogenesis of bony lesions in people with NF1, it is necessary to distinguish primary osseous features from secondary changes. The term dysplasia implies aberrant placement or growth of the cells within a tissue (5). Osseous dysplasia is an intrinsic abnormality of development of bone and would therefore be a primary manifestation of the NF1 mutation. Primary dysplastic lesions differ from secondary bony abnormalities that arise as osseous tissue responds to abnormal mechanical or other local influences. Examples of secondary lesions that may occur in NF1 include bony overgrowth or erosion that develops in response to an adjacent plexiform neurofibroma. This review focuses on the natural history and pathogenesis of the three most frequently occurring and characteristic bony lesions of NF1, those of the vertebrae, long bones, and craniofacial flat bones. We believe that these characteristic osseous lesions are not primary dysplasias, even though they are usually considered to be such in the literature. We believe, rather, that these lesions result from a vicious cycle of abnormal bony maintenance or repair that develops within NF1-haploinsufficient bone in response to abnormal mechanical forces or other distorting influences that affect the bone locally. The contrast between the pathogenic model we propose and the conventional interpretation of these characteristic NF1 bony lesions as dysplasias is summarized in Table 1, and the evidence for our model is described in the following sections. In order to avoid misunderstanding that could be engendered by use of the conventional terms, ‘dysplasia’ and ‘dystrophy’, which imply that the bony lesions characteristic of NF1 are primary features of the disease, we shall use the non-specific term ‘osteopathy’ to describe these lesions.

Craniofacial osteopathy Clinical manifestations

The most distinctive craniofacial osteopathy in NF1 patients is a unilateral defect of the greater wing of the sphenoid bone (12). The lesser wings and the orbital plate of the frontal bone may also be affected (19–21). Some NF1 patients have asymmetry of the facial bones and skull that is not associated with a sphenoid wing lesion (16). Defects of bones along the left lambdoidal and sagittal sutures are also common in NF1 (14). Although sphenoid wing defects may occur in people who do not have NF1, over 50% of cases are associated with NF1. Although sphenoid wing lesions in NF1 are usually asymptomatic, they can easily be diag-

nosed by standard skull radiographs or CT scans. The bony lesions may be associated with elevation of the orbital fissure (22), enlargement of the temporal fossa in all directions, and lifting of the sphenoid ridge out of the orbit (21, 23). There may be direct contact of the temporal lobe of the brain with the orbital soft tissue (16) or, rarely, pulsating exophthalmos (19, 24–27) that can affect vision (21, 25), impair extra-ocular movement, or produce inflammation of the conjunctiva (20). Reported frequencies for sphenoid wing lesions among patients with NF1 range from 3 to 7% (5, 16, 28, 29). Friedman and Birch (6) reported a prevalence of 11.3% among 256 NF1 probands who had skull radiographs, but skull radiographs were only performed in this series for clinical indications such as facial asymmetry, and such patients are more likely to have a sphenoid wing lesion. Although initially thought to be congenital and non-progressive (27), sphenoid wing osteopathy is progressive in some patients with NF1 (30–32). Incomplete formation of the sphenoid bone may be associated with abnormal growth of the skull in children with NF1, producing progressive deformities (30, 32). Progression in adults is much less common and may be indicative of an adjacent tumor that further distorts a deformed sphenoid bone (32). Pathogenesis

The sphenoid bone forms between 4 and 8 developmental weeks in humans and originates from an interaction of mesenchyme, neural crest cells, and neuroectoderm (33). The characteristic sphenoid wing lesions seen in NF1 patients are usually considered to be dysplasias, i.e. primary ossification defects resulting from mesodermal maldevelopment (16, 20, 21, 26) because they are usually congenital and are often not associated with a tumor that is apparent on physical examination or skull radiographs (26). Furthermore, Rovit and Sosman (25) argued that pulsating exophthalmos, which is occasionally seen in association with the sphenoid wing lesions of NF1, is unlikely to be caused by a tumor that replaced the bone because such a tumor would act as a barrier between the orbit and underlying brain and vasculature. However, the fact that sphenoid wing osseous lesion occurs early in life does not necessarily mean that it is an intrinsic abnormality of bony development because diffuse plexiform neurofibromas, which can erode or distort bone secondarily, may develop before birth in NF1. Moreover, this primary dysgenesis hypothesis 379

380

Tibial, vertebral, and sphenoid wing lesions characteristic of NF1

Congenitally abnormal development of only the involved bone(s), i.e. dysplasia of affected bones only

Abnormal mechanical or other distorting forces produced by local factors such as adjacent neurofibromas or vascular lesions induce a vicious cycle of abnormal remodeling and repair because haploinsufficient NF1 bone responds abnormally to distorting forces, i.e. focal osteopathy results from abnormal response of bone to local distorting forces

Examples

Conventional interpretation

Proposed model

Focal osseous lesions

(1) Dysplastic lesions frequently occur in the absence of adjacent neurofibromas or any other distorting influence (2) Dysplastic lesions are all congenital (3) Uninvolved bones respond normally to mechanical and other stresses that develop after birth (4) No abnormality demonstrable in bones that are not involved by dysplastic lesions (1) Osteopathic lesions only develop adjacent to a neurofibroma or other distorting influence (2) Osteopathic lesions may develop either congenitally or later, depending on when abnormal distorting forces occur (3) Osteopathic lesions may spread to involve previously unaffected bone if distorting forces spread from original location (4) Abnormality of osseous maintenance and repair may be demonstrable in bones that are not involved by focal osteopathic lesions

Deficient maintenance and repair produced by NF1 haploinsufficiency

Prediction

Normal

All uninvolved bones in a person with NF1 who has an osseous lesion or all bones in a person with NF1 who has no characteristic osseous lesions

Other bones

Table 1. Comparison of proposed model and conventional interpretation for pathogenesis of focal osseous lesions characteristic of NF1

Alwan et al.

Osseous dysplasia

does not explain the striking predilection for involvement of the sphenoid bone compared to other flat bones of the skull, its usual unilateral occurrence, or its frequent progression over time. Macfarlane et al. (30) argue that destruction of the sphenoid wing is a secondary lesion that results from growth of an orbital plexiform neurofibroma with erosion of the adjacent bony orbit or, alternatively, from increased orbital circulation and expansion of the superior orbital fissure resulting from local vascular abnormality associated with an orbital tumor. Jacquemin et al. (32) performed CT scans on 13 NF1 patients with bony defects of the sphenoid wing and detected a plexiform neurofibroma in the superficial temporal fossa in all cases. The tumor was adjacent to the distorted sphenoid bone in 12 of these cases, and nine also had tumors in the pterygoid fossa. Progression of the sphenoid bone changes was seen in four of these patients who underwent repeated CT scans. These authors propose that the characteristic changes in the sphenoid bone result from interaction with plexiform neurofibromas in utero or early childhood. This hypothesis can explain the progressive nature of sphenoid bone changes and their unilateral occurrence (34). It is interesting to note that the presence of plexiform neurofibromas has also been associated with other changes of the bony orbit, such as enlargement of the orbital rim, decalcification or remodeling of the lateral orbital wall and posterior bony orbit, and enlargement of the orbital foramina (34, 35). Havlik and Boaz (36) invoke an indirect mechanism, attributing sphenoid wing osteopathy to increased cerebrospinal fluid (CSF) pressure. The authors reported a case with unilateral disturbance in CSF dynamics by radionucleotide cisternogram. Indirect evidence of a right-sided increase in CSF pressure included the prominence of the right cerebral hemisphere with a slight midline shift, expansion of the temporal fossa, and herniation of the temporal lobe. This CSF pressure abnormality could be due to early suture closure in the developing skull in response to the presence of an adjacent tumor or abnormal meninges, which occur in many cases of NF1 (32). Another mechanical factor that may of importance in some cases is brain herniation (37, 38). Vertebral osteopathy Clinical manifestations

Riccardi (12) and Akbarnia et al. (39) reported a 10% incidence of scoliosis among NF1 patients. Friedman and Birch (6) found the incidence to be 24%, and Chaglassian et al. (40) estimated that

26% of NF1 patients had scoliosis. Even higher frequencies have been found in some reported series (18, 41, 42). The variability in frequency may be attributable, at least in part, to inconsistency in the definition of scoliosis by different authors and referral bias in these clinic-based studies. About 1% of 10,000 scoliosis cases seen by Winter et al. (43) were found to fit the NF1 diagnostic criteria. NF1 is associated with two different kinds of spinal deformities: ‘dystrophic’ and ‘non-dystrophic’ (16, 43, 44). ‘Dystrophic’ scoliosis is associated with bony abnormalities that are apparent on radiographic examination; ‘non-dystrophic scoliosis usually is not, although vertebral wedging has sometimes been noted (42). Dystrophic scoliosis may or may not involve primary osseous defects, i.e. vertebral ‘dysplasia’ as defined above: some of the bony lesions used to define dystrophic scoliosis are clearly secondary to adjacent plexiform neurofibromas or thickening of adjacent nerves. Moreover, the separation of NF1-associated scoliosis into dystrophic and non-dystrophic curves is made with plain radiographs and may suffer from the inability of such imaging to identify early bony changes (44) as well as adjacent soft tissue lesions. Dystrophic spinal deformities may exhibit severe angulation and rapid progression (12, 15, 43–46). Non-dystrophic deformities, on the other hand, are usually less severe (15, 18, 43). Most NF1 patients with spinal deformities display non-dystrophic scoliosis, which is similar to the idiopathic curvature that is commonly seen in adolescents who do not have NF1 (18, 42). Non-dystrophic scoliosis in NF1 patients is also usually identified in adolescence and affects eight to 10 vertebral segments (42). Management of non-dystrophic scoliosis in NF1 is similar to that of the idiopathic type in patients who do not have NF1. Brace treatment is recommended for curvatures measuring 20–35 degrees; however, the development of pseudarthrosis due to defects in the fusion mass has often been observed after surgical treatment of NF1 patients with ‘non-dystrophic’ curves of more than 35 degrees (42, 44). Progression of non-dystrophic scoliosis to the dystrophic form can also occur in NF1; hence careful follow up is necessary (18, 44). Dystrophic scoliosis in NF1 patients typically produces a progressive sharply angulated curve involving four to six vertebrae in the thoracic area, although shorter or longer segments and other portions of the spine may be involved (16, 18, 40, 43, 44, 46, 47). Kyphoscoliosis may also occur (31). Dystrophic scoliosis usually presents between 6 and 10 years of age and is unlikely to 381

Alwan et al.

develop if it is not apparent by the end of the first decade (12). Dystrophic scoliosis in a child with NF1 may progress rapidly, although progression varies from case to case and is very difficult to predict. The curve and associated bony abnormalities may produce paraparesis or paraplegia (48, 49) and even death (5). In the young child with NF1, the presence of the ‘Riccardi sign’, a hair whorl on the back over the spine, is usually indicative of an underlying neurofibroma and/or vertebral dysplasia and therefore carries a high risk of subsequent development of dystrophic scoliosis (12). Typical NF1 vertebral lesions may include absence or hypoplasia of the vertebral pedicles, posterior scalloping of the vertebrae, or wedging, angulation or rotation of the vertebral bodies (12, 17, 18, 41, 43, 45, 50). The defect may be present in one or a group of vertebrae and varies in severity (12). Vertebral osteopathy is sometimes associated with enlargement of the intervertebral foramina (44). The term ‘vertebral scalloping’ refers to an exaggerated concavity, usually of the posterior side, of the vertebral body (17) (Fig. 1a). Riccardi A

B

Fig. 1. Vertebral osteopathy of neurofibromatosis 1. (a) Computed tomographic image showing vertebral scalloping. (b) A plain chest X-ray demonstrating the appearance of rib penciling.

382

(12) describes posterior vertebral scalloping as highly characteristic of NF1 and the mildest of the associated vertebral defects. He reports it to be present in about 10% of NF1 patients. This defect is not radiographically apparent at birth but becomes evident between the ages of 3 and 10 years (45). The severity of scalloping in affected vertebrae may also progress with age. Kwok et al. (51) assessed lumbar vertebral shape quantitatively in a group of patients with scoliosis with and without NF1. They devised a vertebral ‘scalloping ratio’ and concluded that the distribution of maximum scalloping ratios differed significantly in NF1 patients from others with scoliosis. The extent of wedging and rotation of the vertebral bodies in NF1 patients is also extremely variable (45). Some authors report that vertebral wedging is not associated with curve progression (45), but others conclude that wedging may predict progression of the scoliosis after treatment by joint fusion (52). Rib penciling, a radiographic finding that usually does not produce any symptoms, is characterized by extremely thin ribs with a ‘twisted ribbon’ shape (Fig. 1b) (16, 43, 45, 50). Funasaki et al. (45) found this abnormality in 68% of NF1 patients who had scoliosis. Rib penciling is rarely, if ever, seen in children under 3 years of age, but once it develops, it tends to progress and spread to other thoracic levels (45, 50). Crawford (44, 50) introduced the term ‘modulation’ to describe the ‘ability of a spinal deformity to transform by acquiring various dystrophic morphologic features’. This concept is based on the observation that spinal deformities of NF1, whether initially classified as dystrophic or nondystrophic, may progress and spread to other regions of the spine. Crawford distinguishes ‘modulation’, which implies a change in type from non-dystrophic to dystrophic scoliosis, from ‘progression’, which indicates increased clinical severity whether or not there is an associated change in type. Durrani et al. (50) observed that 81% of spinal deformities that developed before 7 years of age showed evidence of modulation, and the presence of rib penciling was strongly associated with increased progression and modulation. Funasaki et al. (45) observed that early age of onset, severe vertebral rotation, and vertebral scalloping were risk factors for progression of spinal deformities in patients with NF1. Calvert et al. (47) found that curves with anterior vertebral scalloping showed a significantly greater progression than curves without this feature. Neither the severity of other non-skeletal features of NF1 (40) nor the length of the curvature seem to be

Osseous dysplasia

associated with severe progression of dystrophic scoliosis (45, 50). The variable and sometimes rapidly progressive course of dystrophic scoliosis in children with NF1 complicates management. In addition, the frequent occurrence of paravertebral neurofibromas and the occurrence of decreased bone mineral density in patients with NF1 (53) may cause further difficulty. Brace treatment is usually unsuccessful in NF1 patients with dystrophic scoliosis (43, 54, 55). Spinal fusion can be effective but is often technically demanding (46), and some patients with kyphoscoliosis may not attain a solid fusion (40, 43, 47, 56). Laminectomy is not always successful and can cause further deformities (18, 43, 48). Generally, treatment of severe dystrophic scoliosis in NF1 patients remains unsatisfactory, and the post-treatment complications and morbidity are distressingly frequent (40, 46, 56). Pathogenesis

The pathogenesis of vertebral osteopathy in NF1 is not clearly understood, and it remains controversial whether these bony changes represent primary osseous dysplasia or secondary responses to neurofibromas within the intervertebral foramina or adjacent tissues. It is also possible that the pathogenesis differs from case to case. Early investigators (17, 57) suggested that vertebral scalloping in NF1 results from a primary mesodermal dysplasia affecting the bones, dura, and meninges, without involvement of an adjacent neurofibroma. Dural ectasia regularly occurs in association with enlargement of the vertebral foramina and widening of the spinal canal in NF1 patients (17, 58–63), but it is not clear whether dural ectasia is a primary abnormality of the dura or just a response of the dura to changes in the vertebrae and adjacent tissues. Crawford (44) has argued that vertebral scalloping represents bony erosion caused by intraspinal expansion of a neurofibroma, meningocele, or dural ectasia, rather than a primary developmental defect of the bone. It is well known that an expanding neurofibroma can enlarge the intravertebral foramen and erode the pedicle (5, 15, 39) or dorsal portion of a vertebra. MRI and CT studies of patients with vertebral scalloping usually demonstrate an associated paraspinal mass, either a neurofibroma or dural ectasia, or a combination of both (17, 44, 58, 59, 62, 64, 65). Thoracic and lumbar meningoceles have also been associated with vertebral scalloping in some NF1 patients (58, 66–68).

Vertebral wedging in NF1 is widely accepted as a secondary abnormality that results from the pressure of an adjacent paraspinal tumor rather than being a primary osseous defect (18, 45). Similarly, rib penciling appears always to be secondary to either an adjacent neurofibroma or an enlargement of the intercostal nerves (18, 50). Sawatzky et al. (69) characterized the vertebral morphology in nine NF1 patients with dystrophic scoliosis by CT scanning. The spinal canals of the affected vertebrae were widened, with the inner surface of the pedicles and vertebral bodies decreased in size. The pedicles appeared very thin and elongated, which the authors suggest may cause loss of structural integrity and eventual collapse of the spine. Casselman and Mandell (17) speculated that NF1 bone is intrinsically more susceptible to the distorting effects of an adjacent tumor. This could explain why patients with intraspinal tumors who are not affected with NF1 do not exhibit similar dystrophic skeletal abnormalities. More recent observations of decreased bone mineral density in NF1 patients with dystrophic scoliosis and of a correlation of bone mineral density with the severity of kyphoscoliosis (53) are compatible with the possibility that vertebral bone might respond differently to the presence of a spinal tumor in people with NF1 than in others. Long-bone osteopathy Clinical manifestations

Although tibial pseudarthrosis and related lesions are rare in the general population, occurring at a rate of about one in 140,000 (18, 42), the prevalence in people with NF1 is 1–4% (6, 12, 18, 42). Again, the variability of these estimates probably reflects differences in diagnostic criteria and ascertainment biases in these clinic-based series. The tibia is the long bone that is most often involved (18, 70–75), but many NF1 patients have involvement of the fibula of the same leg as well (76). Similar osteopathic lesions have also been reported in the femur (16), radius (77–79), ulna (79), humerus (80), and clavicle (81). Tibial osteopathy usually presents as unilateral anterolateral bowing of the leg, with various degrees of severity (12) (Fig. 2). The term ‘congenital tibial bowing’ is frequently used to describe this condition, although it is usually not recognized in the immediate newborn period. These tibial lesions are usually identified in infancy or early childhood and may precede the appearance of a diagnostic number of cafe´-au-lait spots in children with NF1 (15, 71, 73). 383

Alwan et al.

Fig. 2. Long-bone osteopathy. A radiographic image of the long bones of a neurofibromatosis 1 (NF1) patient showing tibial bowing characteristic of NF1.

Pathological fracture of the affected bone often occurs in the first decade of life and frequently leads to the development of a pseudarthrosis or ‘false joint’ (42, 73, 82). There is inconsistency in the literature with regard to nomenclature of this condition. Some authors refer to all degrees of tibial osteopathy, including anterolateral bowing, as ‘congenital pseudarthrosis’. This is misleading as very few NF1 patients with tibial osteopathy actually have pseudarthrosis at birth, and some never develop pseudarthrosis (12). Pseudarthrosis only occurs as a complication (i.e. non-union) of a frank fracture of the deformed bone (73). Friedman and Birch (6) found pseudarthrosis in 2.2% of 1479 NF1 probands. Forty percent of 340 patients with pseudarthrosis of the tibia also had NF1 in a large multicenter study conducted by the European Pediatric Orthopedic Society (83). Other authors report NF1 in 40 (84), 50 (18), 55 (15, 71, 85), 70 (75), and around 80% (70) of tibial pseudarthrosis patients in smaller clinical series. This variation in the reported prevalence of NF1 among patients with tibial pseudarthrosis has been attributed to the appearance of other 384

NF1 features later in childhood than tibial lesions and to the consequent difficulty of diagnosis of NF1 in young children (41). Although Crawford and Bagamery (15) found an equal sex distribution in NF1 patients with tibial pseudarthrosis, Stevenson et al. (73) and Szudek et al. (86) found a statistically significant male predominance of NF1 cases with long-bone pseudarthrosis. This gender difference was not seen for long-bone osteopathy more generally. Radiographically, affected bones may exhibit cystic lesions, sclerosis, and/or thinning of the cortices (41, 70, 75, 82, 83, 87). Classification systems based on the pathological and radiographic patterns have been proposed by Andersen (70), Boyd (82), and Crawford (15, 42) to provide a basis for prognosis and treatment. Management of pseudarthrosis in NF 1 patients is often difficult because of frequent recurrence of fractures after treatment (41, 42, 82). In patients who are recognized as having tibial defects but have not yet suffered a pathological fracture, bracing the leg until skeletal maturity, when fracture is much less likely to occur, may reduce the chance of developing a pseudarthrosis (42, 54, 82). Once a pseudarthrosis has occurred, treatment options include application of electromagnetic fields, especially when the lesion is associated with a cyst (41, 88), surgical bone grafting (70, 85, 89, 90), vascularized transplant graft using the fibula from the opposite limb (41, 89, 91), and compression and distraction histogenesis of bone and soft tissue with the Ilizarov method (74). No single approach has been found to be effective in all cases, so management must be individualized for each patient. Failure to achieve bony union after repeated attempts at treatment or progression of the lesion may leave amputation as the only option (92). Pathogenesis

The pathogenetic mechanisms involved in tibial osteopathy and the subsequent development of pseudarthrosis are poorly understood. Histopathological studies have been done almost exclusively on tissue obtained from pseudoarthroses, and very little is known about the pathological nature of earlier osseous lesions. Some, but not most, tibial pseudarthroses are associated with the presence of an intraosseous neurofibroma (83, 93). This observation led to the hypothesis that long-bone involvement is not a primary osseous defect but rather a secondary response of the bony tissue to extrinsic pressure (15, 94). The presence of this abnormal tissue within the bone

Osseous dysplasia

may also interfere with healing, thereby causing pseudarthrosis. Electron microscopic examination of the soft tissue associated with tibial pseudarthroses in three NF1 patients revealed the presence of fibroblasts but not Schwann cells, perineurial cells, or axons, which would also be expected in a neurofibroma (95). Similarly, Ippolito et al. (75) reviewed 24 patients, including 17 with NF1, who had long-bone pseudarthrosis and found that the associated histopathological change was the presence of highly cellular fibrovascular tissue that did not appear to be a neurofibroma. No significant differences were revealed in the histopathology of these samples between patients with or without NF1. These authors suggested that this abnormal tissue may interfere with repair of the bony cortex at the fracture site and consequently lead to sclerosis of the medullary canal. Growth of this same tissue in the presence of excessive bone resorption may lead to cystic changes (75). Abdel-Wanis and Kawahara (96) suggested that the abnormal fibrous tissue found in association with tibial pseudarthrosis might result from an abnormal response of periosteal fibroblasts to injury. Molecular characterization of the lesions associated with tibial pseudarthrosis and earlier stages of this osteopathy is needed to understand the mechanisms involved. The NF1 gene and protein

The NF1 gene maps to chromosome 17q11 and is relatively large, covering 300 kb of genomic DNA and comprising 60 exons. It is highly pleiotropic and is expressed in a diverse set of tissues and organs (12, 97). The NF1 protein, neurofibromin, includes a domain that functions to activate ras-GTPase, thereby reducing the activity of an important signal transduction pathway that is involved in the control of cellular growth and differentiation (98). NF1 acts as a tumor-suppressor gene, with some neurofibromas and other NF1associated tumors resulting from complete loss of neurofibromin activity through ‘second-hit’ somatic mutations of the normal NF1 allele (99). Most constitutional mutations in the NF1 gene are nonsense or frameshift mutations that lead to premature truncation of neurofibromin (99, 100). No correlation of the NF1 genotype to phenotype has been recognized (97) except for microdeletions that involve the entire NF1 locus and several adjacent genes. NF1 microdeletions are associated with severe learning disabilities, large numbers of cutaneous neurofibromas that begin to appear at an unusually young age, and an increased

risk of developing malignant peripheral nerve sheath tumors (101–105). The basis for the extreme phenotypic variability that characterizes NF1 as a disease is unknown. Despite of this variability, some individuals affected with NF1 are more likely to develop certain clinical features than others (86, 106–108). Riccardi (5) reported a negative association between the number of cafe´-au-lait spots in an individual with NF1 and the occurrence of tibial pseudarthrosis. A strong association between the occurrence of long-bone pseudarthrosis and the presence of neoplasms other than neurofibromas and optic gliomas was observed in NF1 patients by Szudek et al. (86). They suggest that this association might reflect an overlap of pathogenic mechanisms between neoplasms and pseudarthrosis. Alwan et al. (unpublished data) studied 3377 NF1 probands from the NNFF International Database and found significant associations between the occurrence of sphenoid wing changes and both long bone and vertebral osteopathy, suggesting a common pathogenetic mechanism for development of these characteristic osseous manifestations. Several different factors, including allelic heterogeneity, ‘second-hit’ mutations, somatic mosaicism, epigenetic effects such as imprinting, posttranscriptional control of the NF1 protein, and environmental factors, have been postulated to contribute to the clinical variability of NF1 (97, 109, 110) No information is available regarding the importance of any of these factors in NF1associated osteopathy, but there is some evidence for the involvement of modifying genes. The finding of a higher frequency of long-bone pseudarthrosis in males than females with NF1 (73, 86) is indicative of the role of modifying genes that are directly or indirectly related to gender. Furthermore, Easton et al. (111) found significant clustering of optic gliomas, seizures, and learning disabilities with scoliosis in 175 individuals with NF1 from 48 families and suggested a role of trait-specific modifying genes. The role of neurofibromin in the development of bones is poorly understood. Osteoblasts, or bone-forming cells, and osteoclasts, which are responsible for bone resorption, are both of mesodermal origin. Defects in the activity of either or both of these cell types could be involved in NF1 osteopathy (112). Neurofibromin expression has not been studied in either osteoblasts or osteoclasts, but the NF1 gene is expressed in fibroblasts (113), which are very similar in morphology and gene expression pattern to osteoblasts (96, 112). In addition, neurofibromin is expressed in embryonic chondrocytes, which 385

Alwan et al.

produce the cartilaginous template for bone development (114). Both osteoblastic and osteoclastic cells play an important role in the regulation of bone mass (115). Osteomalacia, pathological softening of bone caused by inadequate mineralization of osteoid, occurs more often than expected among people with NF1 (16, 116). The condition typically presents in mid-adulthood with bone pain, multiple pseudofractures, hypophosphatemia, and renal phosphate loss. The pathophysiology of this form of osteomalacia is unclear. However, mildly decreased bone mineral density appears to be very common among NF1 patients. Illes et al. (53) found decreased bone mineral density in the vertebrae of NF1 patients who have a spinal deformity, and a more recent investigation (117) found both lumbar bone mineral density and whole-body bone mineral content to be significantly reduced among NF1 patients. Discussion

Any unifying hypothesis for development of the characteristic osseous lesions in NF1 needs to account for the following observations: (i) their spotty (unilateral, involvement of some vertebrae but not others) distribution despite the fact that the constitutional NF1 mutation presumably affects all cells of the body; (ii) their obvious predilection for only a few bones – especially the vertebrae, sphenoid wing, tibia, and fibula; (iii) their occurrence in only a minority of NF1 patients and (rarely) in patients who do not have NF1; (iv) the onset at birth or early in childhood; (v) the progressive nature of many of the lesions; and (vi) the occurrence of clinically indistinguishable lesions in association with adjacent neurofibromas or other soft tissue lesions in many cases. It is clear that the characteristic osseous lesions of NF1 do not all arise as primary abnormalities of bony development, i.e. as dysplasias in the strict sense. However, we do not know whether the presence of an abnormality (e.g. an adjacent neurofibroma) that distorts the bone locally is necessary or just facilitative for the development of NF1 osteopathy. Similarly, we do not know if an intrinsic abnormality of the bone is necessary, and, if so, whether all bones in people with NF1 possess the same potential to develop these lesions or not. For example, the characteristic tibial osteopathy might arise as a primary dysplasia that results from a ‘second-hit’ somatic mutation of the normal NF1 allele in the anlage of the affected bone. If this were the case, however, it is hard to 386

understand why the tibia would be involved so much more often than any other long bone. Alternatively, all NF1 bone might be subtly abnormal in its capacity for maintenance and repair, but typical NF1 osteopathy only develops when other pathogenic factors is also present. If this were true, the unusual distribution of typical skeletal lesions might reflect uncommon additional pathogenic factors such as a plexiform neurofibroma within a closed space like the orbit or the extraordinary mechanical stresses that might occur in the anteriorly bowed tibia of a toddler or in the scoliotic spine of a child. We believe that all of the known clinical, radiological, and pathological features of NF1 osteopathies can be explained by the pathogenic model shown in Fig. 3 and summarized in Table 1. We hypothesize that the characteristic osseous lesions arise when NF1-haploinsufficient bone responds abnormally to mechanical stimuli that normally produce efficient remodeling of bone. In a person with NF1, this abnormal osseous remodeling may set up a vicious cycle because the new NF1haploinsufficient bone is mechanically poorly adapted, stimulating more inefficient osseous

Normal

Abnormal biomechanical forces

NF1

Haploinsufficient bone with adjacent neurofibroma Fig. 3. Model for development of typical osseous lesions in neurofibromatosis 1 (NF1). In normal bone, normal process of remodelling produces a strong and mechanically well-adapted bone. NF1-haploinsufficient bone remodels less effectively in response to normal biomechanical stresses. When bone is further compromised by the presence of an adjacent neurofibroma, inefficient remodelling is exacerbated, producing more abnormal, mechanically poorly adapted bone, which stimulates more inefficient remodelling by NF1-haploinsufficient bone.

Osseous dysplasia

remodeling, and so on. Neither the intrinsic abnormality that affects all NF1-haploinsufficient bone nor the distorting mechanical or other factor that triggers the vicious cycle is sufficient by itself to produce the characteristic NF1 osteopathy; both are necessary. One question that arises is what the initiating event for this process is in a particular bone. With vertebral or sphenoid wing abnormalities, an adjacent neurofibroma that erodes bone is usually responsible. In other circumstances, no neurofibroma is present, and overgrowth of another tissue, hypervascularity, an incidental mechanical deformation (e.g. uterine constraint of the leg), or some other factor may be involved. Critical evaluation of this hypothesis requires a better understanding of the natural history of the characteristic osseous lesions of NF1. Unfortunately, most available investigations are limited in size, include pathology studies only of endstage lesions, and/or employ only plain radiography to image the lesions. Although more recent CT and MRI studies are limited, they are consistent with the proposed model. For example, CT and MRI studies of sphenoid osteopathy have consistently demonstrated the presence of associated tumors, often in direct contact with the osseous defect (30, 32, 36). Increased pressure on bone from CSF, expansion of the dural sac, and pulsating exophthalmos may also be involved in some cases. The striking similarity between the expansion of the temporal fossa and resorption or thinning of the sphenoid bone that occurs in NF1 and in cases of temporal arachnoid cyst or chronic subdural hematoma adds further evidence to the contribution of increased intracranial pressure as an underlying mechanism (36). Similarly, this review has presented evidence of secondary destruction of the vertebrae by contiguous intraspinal tumors, dural ectasia, or meningocele (58, 61–68) and of long-bone osteopathy being commonly associated with the presence of abnormal pathological tissue (75, 83). The fact that patients with intraspinal pathologies, who do not have NF1, do not usually develop vertebral osteopathy characteristic of NF1 suggests that an ‘intrinsic abnormality’ of NF1 bone is also important pathogenically. The occurrence of a generalized decrease in bone mineral density (117) and the occasional development of frank osteomalacia among people with NF1 are consistent with this interpretation. An intrinsic abnormality of bone remodeling might also explain the poor bone graft resorption that has been described following orbital reconstruction (36), the frequent recurrence of fracture after treatment of tibial pseudarthrosis (42), the

failure of fusion and instrumentation that may occur after surgical treatment of severe scoliosis (40, 42–44, 47, 56), and the progressive nature of many of the typical osseous lesions of NF1. CT and MRI studies are needed on larger numbers of patients with NF1 osteopathy to confirm the pathogenetic mechanisms involved. Expression studies of the NF1 gene and protein in apparently healthy and diseased tissue are also needed. The development of an animal model of NF1 osteopathy would be very useful. Improved understanding of the pathogenesis of NF1 osteopathy should permit better management of affected patients. Recognition of adjacent soft tissue pathologies with craniofacial, vertebral, and long-bone lesions should help identify patients who are at high risk of developing debilitating osseous abnormalities and provide the opportunity for earlier and more effective treatment. Acknowledgements We would like to extend special thanks to Manraj KS Heran, MD, FRCPC, Department of Radiology, University of British Columbia, for providing us with some of the radiographic images for this review.

References 1. NIH. Neurofibromatosis: Conference Statement. National Institutes of Health Consensus Development Conference. Arch Neurol 1988: 45: 575–578. 2. Friedman JM. Epidemiology of neurofibromatosis type 1. Am J Med Genet 1999: 89: 1–6. 3. Rasmussen SA, Yang Q, Friedman JM. Mortality in neurofibromatosis 1: an analysis using U.S. death certificates. Am J Hum Genet 2001: 68: 1110–1118. 4. Gutmann DH, Aylsworth A, Carey JC et al. The diagnostic evaluation and multidisciplinary management of neurofibromatosis 1 and neurofibromatosis 2. JAMA 1997: 278: 51–57. 5. Riccardi VM. Neurofibromatosis: phenotype, natural history, and pathogenesis, 2nd edn. Baltimore: John Hopkins University Press, 1992. 6. Friedman JM, Birch PH. Type 1 neurofibromatosis: a descriptive analysis of the disorder in 1,728 Patients. Am J Med Genet 1997: 70: 138–143. 7. DeBella K, Szudek J, Friedman JM. Use of the National Institutes of Health Criteria for diagnosis of neurofibromatosis 1 in children. Pediatrics 2000: 105: 608–614. 8. Bolande RP. Neurofibromatosis – the quintessential neurocristopathy. pathogenetic concepts and relationships. Adv Neurol 1981: 29: 67–75. 9. Clementi M, Milani S, Mammi I et al. Neurofibromatosis type 1 growth charts. Am J Med Genet 1999: 87: 317–323. 10. Szudek J, Birch P, Friedman JM. Growth in North American white children with neurofibromatosis 1 (NF1). J Med Genet 2000: 37: 933–938. 11. Szudek J, Birch P, Friedman JM. Growth charts for young children with neurofibromatosis 1 (NF1). Am J Med Genet 2000: 92: 224–227.

387

Alwan et al. 12. Riccardi VM. Skeletal system. In: Neurofibromatosis: Phenotype, Natural History and Pathogenesis (Friedman J, Gutmann D, MacCollin M, Riccardi VM, eds), 3rd edn. Baltimore: John Hopkins University Press, 1999: 250–273. 13. Brooks B, Leham EP. The bone changes in Recklinghausen’s neurofibromatosis. Surg Gynecol Obstet 1924: 38: 587–595. 14. Holt JF. Neurofibromatosis in children. AJR AM J Roentgenol 1978: 130: 615–639. 15. Crawford AH, Bagamery N. Osseous manifestations of neurofibromatosis in childhood. J Pediatr Orthop 1986: 6: 72–88. 16. Hunt JC, Pugh DG. Skeletal lesions in neurofibromatosis. Radiology 1961: 76: 1–19. 17. Casselman ES, Mandell GA. Vertebral scalloping in neurofibromatosis. Radiology 1979: 131: 89–94. 18. Vitale MG, Guha A, Skaggs DL. Orthopedic manifestations of neurofibromatosis in children. An Update. Clin Orthop 2002: 401: 107–118. 19. Wiesenfeld D, James PL. Pulsating exophthalmos associated with neurofibromatosis. J Maxillofac Surg 1984: 12: 11–13. 20. Myrianthopoulos NC. Other skull defects: craniolacunia, foramina parietalia permagna, aplasia of the sphenoid wings. In: Handbook of Clinical Neurology (Vinken PJ, Bruyn GW, eds), 30, Amsterdam: Elsevier, 1977: 269–284. 21. Burrows EH. Bone changes in orbital neurofibromatosis. Br J Radiol 1963: 36: 549–561. 22. Harkens K, Dolan KD. Correlative imaging of sphenoid dysplasia accompanying neurofibromatosis. Ann Otol Laryngol 1990: 99: 137–141. 23. Zimmerman RA, Bilaniuk LT, Metzger RA et al. Computed tomography of orbitalfacial neurofibromatosis. Radiology 1983: 146: 113–116. 24. Clauser L, Carinci F, Galie M. Neurofibromatosis of the orbit and skull base. J Craniofac Surg 1998: 9: 280–284. 25. Rovit RL, Sosman MC. Hemicrania aplasia with pulsating exophthalmos. An unusual manifestation of von Recklinghausen’s disease. J Neurosurg 1960: 17: 104–121. 26. Binet EF, Kieffer SA, Martin SH et al. Orbital dysplasia in neurofibromatosis. Radiology 1969: 93: 829–833. 27. Rao SB, Dinakar I. Congenital absence of the sphenoid wing. Indian J Radiol 1970: 24: 282–284. 28. Bognanno JR, Edwards MK, Lee TA et al. Cranial MR imaging in neurofibromatosis. AJR Am J Roentgenol 1988: 151: 381–388. 29. White AK, Smith RJ, Bigler CR et al. Head and neck manifestations of neurofibromatosis. Laryngoscope 1986: 96: 732–737. 30. Macfarlane R, Levin AV, Weksberg R et al. Absence of the greater sphenoid wing in neurofibromatosis type 1: congenital or acquired: case report. Neurosurgery 1995: 37: 129–133. 31. Miller A. Neurofibromatosis with reference to skeletal changes. Compression myelitis and malignant degeneration. Arch Surg 1936: 32: 109–122. 32. Jacquemin C, Bosely TM, Liu D et al. Reassessment of sphenoid dysplasia associated with neurofibromatosis Type1. AJNR Am J Neuroradiol 2002: 23: 644–648. 33. Jacquemin C, Mullaney P, Bosley TM. Abnormal development of the lesser wing of the sphenoid with microphthalmos and microcephaly. Neuroradiology 2001: 43: 178–182. 34. Jacquemin C, Bosely TM, Svedberg H. Orbit deformities in craniofacial neurofibromatosis type 1. AJNR Am J Neuroradiol 2003: 24: 1678–1682.

388

35. Reed D, Robertson WD, Rootman J, Douglas G. Plexiform neurofibromatosis of the orbit. CT evaluation. AJNR Am J Neuroradiol 1986: 7: 259–263. 36. Havlik RJ, Boaz J. Cranio-orbital-temporal neurofibromatosis: are we treating the whole problem? J Craniofac Surg 1998: 9: 529–535. 37. Schut L. Absence of the greater sphenoid wing in neurofibromatosis type 1: congenital or acquired: case report. Neurosurgery 1995: 37: 133 [Comment]. 38. Ciricillo SF, Edwards MSB. Absence of the greater sphenoid wing in neurofibromatosis type 1: congenital or acquired: case report. Neurosurg 1995: 37: 133 [Comment]. 39. Akbarnia BA, Gabriel KR, Beckman E et al. Prevalence of scoliosis in neurofibromatosis. Spine 1992: 17: S244–S248. 40. Chaglassian JH, Riseborough EJ, Hall JE. Neurofibromatous scoliosis. Natural history and results of treatment in thirty seven cases. J Bone Joint Surg 1976: 58: 695–702. 41. Crawford AH. Neurofibromatosis in children. Acta Orthop Scand Suppl 1986: 218: 1–60. 42. Crawford AH, Schorry EK. Neurofibromatosis in children: the role of the orthopaedist. J Am Acad Orthop Surg 1999: 7: 217–230. 43. Winter RB, Moe JH, Bradford DS et al. Spine deformity in neurofibromatosis. J Bone Joint Surg 1979: 61: 677–694. 44. Crawford AH. Pitfalls of spinal deformities associated with neurofibromatosis in children. Clin Orthop 1989: 245: 29–42. 45. Funasaki H, Winter RB, Lonstein JB et al. Pathophysiology of spinal deformities in neurofibromatosis. An analysis of seventy-one patients who had curves associated with dystrophic changes. J Bone Joint Surg Am 1994: 76: 692–700. 46. Sirois JL, Drennan JC. Dystrophic spinal deformity in neurofibromatosis. J Pediatr Orthop 1990: 10: 522–526. 47. Calvert PT, Edgar MA, Webb PJ. Scoliosis in neurofibromatosis. The natural history with and without operation. J Bone Joint Surg 1989: 71: 246–251. 48. Winter RB, Lonstein JE, Anderson M. Neurofibromatosis hyperkyphosis. A review of 33 patients with kyphosis of 80 or greater. J Spinal Disord 1988: 1: 39–49. 49. Curtis BH, Risher RL, Butterfield WL et al. Neurofibromatosis with paraplegia: Report of eight cases. J Bone Joint Surg 1969: 51: 843–861. 50. Durrani AA, Crawford AH, Chouhdry SN et al. Modulation of spinal deformities in patients with neurofibromatosis type 1. Spine 2000: 25: 69–75. 51. Kwok ESH, Sawatzky B, Birch P et al. Vertebral scalloping in neurofibromatosis type 1: a quantitative approach. Can J Surg 2002: 45: 181–184. 52. Wilde PH, Upadhyay SS, Leong JC. Deterioration of operative correction in dystrophic spinal neurofibromatosis. Spine 1994: 19: 1264–1270. 53. Illes T, Halmai V, de Jonge T, Dubousset D. Decreased bone mineral density in neurofibromatosis-1 patients with spinal deformities. Osteoporos Int 2001: 12: 823–827. 54. DiSimone RE, Berman AT, Schwentker EP. The orthopedic manifestations of neurofibromatosis. A clinical experience and review of the literature. Clin Orthop 1988: 230: 277–283. 55. Kim HW, Weinstein SL. Spine update. The management of scoliosis in neurofibromatosis. Spine 1997: 22: 2770–2776. 56. Hsu LC, Lee PC, Leong JC. Dystrophic spinal deformities in neurofibromatosis. Treatment by anterior and posterior fusion. J Bone Joint Surg Br 1984: 66: 495–499. 57. Salerno NR, Edeiken J. Vertebral scalloping in neurofibromatosis. Radiology 1970: 97: 509–510.

Osseous dysplasia 58. Egelhoff JC, Bates DJ, Ross JS et al. Spinal MR findings in neurofibromatosis types 1 and 2. AJNR 1992: 13: 1071–1077. 59. Klatte EC, Franken EA, Smith JA. The radiographic spectrum in neurofibromatosis. Semin Roentgenol 1967: 2: 17–33. 60. Mitchell GE, Lourie H, Berne AS. The various causes of scalloped vertebrae with notes on their pathogenesis. Radiology 1967: 89: 67–74. 61. Bensaid AH, Dietemann JL, Kastler B et al. Neurofibromatosis with dural ectasia and bilateral symmetrical pedicular clefts: report of two cases. Neuroradiology 1992: 34: 107–109. 62. Jamjoom ZAB, Malabarey T, Kolawole TM, Araby K. Lumbosacral dural anomalies in Von Recklinghausen neurofibromatosis. Acta Neurochir 1990: 105: 140–146. 63. Stone JW, Bridwell KH. Dural ectasia associated with spontaneous dislocation of the upper part of the thoracic spine in neurofibromatosis. J Bone Joint Surg 1987: 69: 1079–1083. 64. Schonauer C, Tessitore E, Frascadore L et al. Lumbosacral dural ectasia in type 1 neurofibromatosis. Report of two cases. J Neurosurg Sci 2000: 44: 165–168. 65. Tubbs RS, Oakes WJ. Dura ectasia in neurofibromatosis. Pediatr Neurosurg 2002: 37: 331–332. 66. Drevelengas A, Kalaitzoglou I. Giant lumbar meningocele in a patient with neurofibromatosis. Neuroradiology 1995: 37: 195–197. 67. Rainov NG, Heidecke V, Burkert W. Thoracic and lumbar meningocele in neurofibromatosis type 1. Report of two cases and review of the literature. Neurosurg Rev 1995: 18: 127–134. 68. Nakasu Y, Minouchi K, Hatsuda N. Thoracic meningocele in neurofibromatosis. CT and MR findings. J Comput Assist Tomogr 1991: 15: 1062–1064. 69. Sawatzky BJ, Tredwell SJ, Majoor JWM. Vertebral morphology of the dystrophic spine in neurofibromatosis 1. 58th Annual Meeting of the Canadian Orthopaedic Association. Winnipeg, Manitoba October 3–5, 2003. 70. Andersen KS. Congenital pseudarthrosis of the Leg. (Late Results). J Bone Joint Surg Am 1976: 58A: 657–662. 71. Crawford AH. Neurofibromatosis in the pediatric patient. Symp Pediatr Orthop 1978: 9: 11–23. 72. Dunn AW. Case of overgrowth of leg and anterolateral bowing of tibia in neurofibromatosis. Am J Orthop 1965: 7: 120–123. 73. Stevenson DA, Birch PH, Friedman JM et al. Descriptive analysis of tibial pseudarthrosis in patients with neurofibromatosis 1. Am J Med Genet 1999: 84: 413–419. 74. Boero S, Catagni M, Donzelli O et al. Congenital pseudarthrosis of the tibia associated with neurofibromatosis 1: treatment with Ilizarov’s device. J Pediatr Orthop 1997: 17: 675–684. 75. Ippolito E, Corsi A, Grill F et al. Pathology of bone lesions associated with congenital pseudarthrosis of the leg. J Pediatr Orthop B 2000: 9: 3–10. 76. Keret D, Bollini G, Dungl P et al. The fibula in congenital pseudarthrosis of the tibia: The EPOS multicenter study. J Pediatr Orthop B 2000: 9: 69–74. 77. Gregg PJ, Price BA, Ellis HA et al. Pseudarthrosis of the radius associated with neurofibromatosis. Clin Orthop 1982: 171: 175–179. 78. Masihuz-Zaman. Pseudarthrosis of the radius associated with neurofibromatosis. J Bone Joint Surg 1977: 59: 977–978. 79. Ali MS, Hooper G. Congenital pseudarthrosis of the ulna due to neurofibromatosis. J Bone Joint Surg 1982: 64: 600–602.

80. Floyd A, Percy-Lancaster R. The elephant woman. Neurofibromatosis associated with pseudarthrosis of the humerus. J Bone Joint Surg Br 1987: 69: 121–123. 81. Alldred AJ. Congenital pseudarthrosis of the clavicle. J Bone Joint Surg 1963: 45-B: 312–319. 82. Boyd HB. Pathology and natural history of congenital pseudarthrosis of the tibia. Clin Orthop 1982: 166: 5–13. 83. Hefti F, Bollini G, Dungl P et al. Congenital pseudarthrosis of the tibia: history, etiology, classification, and epidemiological data. J Pediatr Orthop B 2000: 9: 11–15. 84. Sofield HA. Congenital pseudarthrosis of the tibia. Clin Orthop 1971: 76: 33–42. 85. Hardinge K. Congenital anterior bowing of the tibia. The significance of the different types in relation to pseudarthrosis. Ann R Coll Surg Engl 1972: 51: 17–30. 86. Szudek J, Evans DG, Friedman JM. Patterns of associations of clinical features in neurofibromatosis 1 (NF1). Hum Genet 2003: 112: 289–297. 87. Cheema J, Grissom LE, Harcke HT. Radiographic characteristics of lower-extremity bowing in children. Radiographics 2003: 23: 871–880. 88. Kort JS, Schink MM, Mitchell SN et al. Congenital pseudarthrosis of the tibia: Treatment with pulsing electromagnetic fields – The international experience. Clin Orthop 1982: 165: 124–127. 89. Weiland AJ, Weiss APC, Moore JR, Tolo VT. Vascularized fibular grafts in the treatment of congenital pseudarthrosis of the tibia. J Bone Joint Surg Am 1990: 72: 654–662. 90. Coleman SS, Coleman DD. Treatment by transfer of the ipsilateral fibula with vascular pedicle. J Pediatr Orthop 1994: 14: 156–160. 91. Gordon L, Weulker N, Jergesen H. Vascularized fibular grafting for the treatment of congenital pseudarthrosis of the tibia. Orthopedics 1986: 9: 825–831. 92. McCarthy RE. Amputation for congenital pseudarthrosis of the tibia. Indications and Techniques. Clin Orthop 1982: 166: 58–61. 93. Green WT, Rudo N. Pseudarthrosis and neurofibromatosis. Arch Surg 1943: 46: 639–651. 94. Aegerter EE. The possible relationship of neurofibromatosis, congenital pseudarthrosis and fibrous dysplasia. J Bone Joint Surg 1950: 32: 618–626. 95. Briner J, Yunis E. Ultrastructure of congenital pseudarthrosis of the tibia. Arch Pathol 1973: 95: 97–99. 96. Abdel-Wanis ME, Kawahara N. Bone development in neurofibromatosis 1. Med Hypotheses 2003: 60: 459–462. 97. Carey JC, Viskochil DH. Neurofibromatosis type 1. A model condition for the study of the molecular basis of variable expressivity in human disorders. Am J Med Genet 1999: 89: 7–13. 98. Xu GF, O’Connell P, Viskochil D et al. The neurofibromatosis type 1 gene encodes a protein related to GAP. Cell 1990: 62: 599–608. 99. Von-Deimling A, Krone W, Menon AG. Neurofibromatosis type 1: pathology, clinical features and molecular genetics. Brain Pathol 1995: 5: 153–162. 100. Viskochil DH. The structure and function of the NF1 gene: molecular pathophysiology. In: Neurofibromatosis: phenotype, natural history and pathogenesis (Friedman J, Gutmann D, MacCollin M, Riccardi VM, eds), 3rd edn. Baltimore: John Hopkins University Press, 1999: 119–141. 101. Kayes LM, Riccardi VM, Burke W et al. Large de novo DNA deletion in a patient with sporadic neurofibromatosis 1, mental retardation, and dysmorphism. J Med Genet 1992: 29: 686–690.

389

Alwan et al. 102. Leppig KA, Kaplan P, Viskochil D et al. Familial neurofibromatosis 1 microdeletions: cosegregation with distinct facial phenotype and early onset of cutaneous neurofibromata. Am J Med Genet 1997: 73: 197–204. 103. Tonsgard JH, Yelavarthi KK, Cushner S et al. Do NF1 gene deletions result in a characteristic phenotype? Am J Med Genet 1997: 73: 80–86. 104. Wu B-L, Schneider GH, Korf BR. Deletion of the entire NF1 gene causing distinct manifestations in a family. Am J Med Genet 1997: 69: 98–101. 105. Kluwe L, Friedrich RE, Peiper M et al. Constitutional NF1 mutations in neurofibromatosis 1 patients with malignant peripheral nerve sheath tumors. Hum Mutat 2003: 22: 420. 106. Szudek J, Joe H, Friedman JM. Analysis of intrafamilial phenotypic variation in neurofibromatosis 1 (NF1). Genet Epidemiol 2002: 23: 150–164. 107. Szudek J, Birch P, Riccardi VM et al. Associations of clinical features in neurofibromatosis 1 (NF1). Genet Epidemiol 2000: 19: 429–439. 108. Szudek J, Friedman JM. Unidentified bright objects associated with features of neurofibromatosis 1. Pediatr Neurol 2002: 27: 123–127. 109. Riccardi VM. Genotype, malleotype, phenotype, and randomness: lessons from neurofibromatosis-1 (NF-1). Am J Hum Genet 1993: 53: 301–304. 110. Friedman JM, Riccardi VM. Clinical and epidemiological features. In: Neurofibromatosis: phenotype, natural

390

111.

112.

113.

114.

115. 116.

117.

history and pathogenesis (Friedman J, Gutmann D, MacCollin M, Riccardi VM, eds), 3rd edn. Baltimore: John Hopkins University Press, 1999: 29–86. Easton DF, Ponder MA, Huson SM et al. An analysis of variation in expression of neurofibromatosis (NF) type 1 (NF1): evidence for modifying genes. Am J Hum Genet 1993: 53: 305–315. Ducy P, Schinke T, Karsenty G. The osteoblast: a sophisticated fibroblast under central surveillance. Science 2000: 289: 1501–1504. Atit RP, Crowe MJ, Greenhalgh DG et al. The Nf1 tumor suppressor regulates mouse skin wound healing, fibroblast, proliferation and collagen deposited by fibroblasts. J Invest Dermatol 1999: 112: 835–842. Daston MM, Ratner N. Neurofibromin, a predominantly neuronal GTPase activating protein in the adult, is ubiquitously expressed during development. Dev Dyn 1992: 195: 216–226. Teitelbaum SL. Bone resorption by osteoclasts. Science 2000: 289: 1504–1508. Konishi K, Nakamura M, Yamakawa H et al. Hypophosphatemic osteomalacia in von Recklinghausen neurofibromatosis. Am J Med Sci 1991: 301: 322–328. Brunnetti-Pierri N, Phan K, Carter S et al. Generalized osteopenia in neurofibromatosis type 1 patients points to an underlying disorder of skeletal homeostasis and mineralization. Am J Hum Genet 2003: 73: 261.